Visceral fat, liver, and skeletal muscle are all major target organs of insulin and play key roles in the pathophysiology of type 2 diabetes. Not known is whether visceral fat, liver fat, and thigh skeletal muscle area or fat directly measured by computed tomography (CT) are independently associated with the incidence of type 2 diabetes. In a prospective cohort study, 1228 nondiabetic Japanese men and women were included. Intra-abdominal fat area (IAFA), liver-to-spleen (L/S) attenuation ratio to assess liver fat, and mid-thigh muscle area and attenuation were measured by CT. Low CT attenuation of liver and muscle reflects greater liver and intramuscular lipid content. Insulin resistance was assessed by HOMA-IR. Type 2 diabetes was diagnosed as fasting plasma glucose level ≥126 mg/dL, HbA1c ≥6.5%, or taking oral hypoglycemic medications or insulin. Discrete-time logistic hazard model was used. During the 7686 person-years of follow-up (median follow-up period, 7 years), 95 subjects developed type 2 diabetes. IAFA and L/S ratio, not thigh muscle area, were independently associated with the risk of type 2 diabetes. In multiple-adjusted models that included IAFA, L/S ratio, thigh muscle area, age, gender, family history of diabetes, HOMA-IR, smoking habit, daily alcohol consumption, and regular physical activity, odds ratios of type 2 diabetes for tertile 2 and 3 of IAFA were 2.23 (95% CI, 0.94-5.31) and 3.16 (1.33-7.52), respectively, compared to tertile 1, and those for tertile 1 and 2 of L/S ratio were 2.01 (1.01-3.97) and 2.(1.04-4.08), respectively, compared to tertile 3. However, those of tertile 2 and 3 of thigh muscle area were 0.76 (0.32-1.78) and 0.64 (0.24-1.69), respectively, compared to tertile 1. When thigh muscle attenuation was substituted for its area, results was identical to the above model.

In conclusion, both greater amounts of visceral and liver fat, not thigh muscle area nor attenuation, were independently associated with the risk of type 2 diabetes.


K. Sato: None. S. Uehara: None. M. Shibata: None. Y. Hikita: None. W.Y. Fujimoto: None. E.J. Boyko: None. T. Hayashi: None.

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at